1. Field of the Invention
This invention relates to an image capturing device and image synthesis method thereof and particularly to an image capturing device and image synthesis method thereof that may analyze a shot scene on multiple sorts of exposure brightness to further synthesize a high dynamic range image.
2. Description of the Related Art
A high dynamic range (HDR) file is a graphics file of special format, and each pixel of the file contains normal RGB information and real brightness information. However, each pixel in a normal graphics file is defined to have a gray scale ranged only from 0 to 255, which is actually not enough because the gray scale or brightness range between the light-emitting strength and a real black object is far beyond 256 levels. Thus, when a normal daytime landscape picture are displayed, the white cloud area and the sun area might be shown in the same gray scale/brightness, but the white cloud and the sun actually might be quite different in brightness. Accordingly, the normal graphics file format is not very definite since no real conditions are kept.
The high dynamic range indicates a contrast ratio of a brightest portion to a darkest portion in a scene. In a high dynamic range image, a brightness value of real scene that is fairly beyond 256 levels is recorded, in which a portion beyond the levels is not shown in the display. It is possibly regarded that a picture shot out of a window inside, which is opened in Photoshop, shows an over exposed portion outside the window and the portion stays all white without any details. Dimming cannot solve everything because it brings merely gray and does not show more details. If the same scene is recorded from the high dynamic range image, decreasing the exposure degree makes the pure white portion show more details.
Currently, a synthesized high dynamic range image is mostly made for a scene with a fixed exposure value from several images, in which each image is poor slightly at exposure compensation of 1 or 2 exposure values (EV) or lower, which is tried to combine many different exposed images for the dynamic image regarding the scene. However, in the existing method, the dynamic state of each image is limited, so multiple images must be taken for synthesis; a general camera is limited in processing capability and thus if more than three images cannot be handled in a short time, a few images must be selected for synthesis. For example, 2 or 3 images are processed and constant exposure compensation is used to capture 2 or 3 images for the synthesis of high dynamic range images. However, in this way, sometimes owing to the scene, an appropriate value of exposure compensation cannot effectively selected to capture or handle the images, and thus a desired high dynamic range image often cannot be produced. Therefore, for the demand, designing a perfect image capturing device and image synthesis method to synthesize high dynamic range images has become an urgent issue for the market application.